• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

New Schiit Yggdrasil with TI chip

GXAlan

Major Contributor
Forum Donor
Joined
Jan 15, 2020
Messages
4,038
Likes
6,302

They claim that blind testing favors multibit DACs even though measurements fall below traditional AKM/ESS sigma delta designs.

What is interesting is that this hits pretty good 1 kHz SINAD and not as great 20 kHz SINAD but good for multibit.

1698940683620.png


1698940660837.png


Any thoughts? I mean SINAD at 20 kHz matters less, because we cannot hear 40 kHz right? Is there IMD that folds down into the audible range?
 

gvl

Major Contributor
Joined
Mar 16, 2018
Messages
3,528
Likes
4,129
Location
SoCal
Is this yet another new chip? Google for GoldenSound measurements, they showed some disconcerting artifacts in audible range that resulted from HF stimulus, if memory serves.
 

IAtaman

Major Contributor
Forum Donor
Joined
Mar 29, 2021
Messages
2,461
Likes
4,338

They claim that blind testing favors multibit DACs even though measurements fall below traditional AKM/ESS sigma delta designs.

What is interesting is that this hits pretty good 1 kHz SINAD and not as great 20 kHz SINAD but good for multibit.





Any thoughts? I mean SINAD at 20 kHz matters less, because we cannot hear 40 kHz right? Is there IMD that folds down into the audible range?
1KHz SINAD is A-weigthed I think and is with 20K bandwith wheras the 20KHz test is not weigthed and 90K bandwidth so hard to compare and understand what exactly is the SINAD performance in a way its comparable to other DACs Amir measured.

I think the problem of converting digital representation of audio data to analog voltage has been practically resolved. I understand companies need differentiation but I am not sure this is the way to go. Schiit is in California no, so why not hire some software guys build some in-house SW capability and give us a PEQ , and come up with some specific tunings and crossfades and effects that you have IP protected and sell us on that - not on multibit DAC. Anyway, its just me :)
 
Last edited:

NTK

Major Contributor
Forum Donor
Joined
Aug 11, 2019
Messages
2,779
Likes
6,174
Location
US East

They claim that blind testing favors multibit DACs even though measurements fall below traditional AKM/ESS sigma delta designs.

What is interesting is that this hits pretty good 1 kHz SINAD and not as great 20 kHz SINAD but good for multibit.

View attachment 323229

View attachment 323228

Any thoughts? I mean SINAD at 20 kHz matters less, because we cannot hear 40 kHz right? Is there IMD that folds down into the audible range?
Note that the measurement bandwidth of the top graph (1 kHz test, see the "Filters" parameters in the leftmost pane in the screen shot) is 20 kHz, whereas the bandwidth used for the 20 kHz test is 90 kHz. Not apple-to-apple comparison.

The claimed advantage by TI for this R-2R DAC is that there is no increase in the noise magnitude in the out-of-band frequencies (i.e. flat noise density with frequency), whereas ΣΔ DAC's have high out-of-band noise due to noise shaping. It is an advantage for certain applications, but not for consumer audio reproduction.
 
OP
G

GXAlan

Major Contributor
Forum Donor
Joined
Jan 15, 2020
Messages
4,038
Likes
6,302
why not hire some software guys build some in-house SW capability and give us a PEQ , and come up with some specific tunings and crossfades and effects that you have IP protected and sell us on that - not on multibit DAC. Anyway, its just me :)

I do wonder. I am slower than most but even I have made the leap to active speakers, as have companies like Harman and Klipsch, along with every soundbar being used today.

There used to be satisfaction in having the best, like owning a car with top speeds or highest fording depth even if you didn’t use that feature. Or maybe even a diamond that retains its clarity even under a microscope to use inclusive examples.

At least when it comes to audio, I used to be in the perspective of going for mathematical transparency, season with tubes or DSEE HX as appropriate. But when Genelec gets world class sound with a simple 100 dB SINAD all l-in-one codec, it makes you wonder if any of this really matters?

The audiophile hobby isn't dying because because people spend less time with focused music listening, it’s that there is less gatekeeping and you probably get great results with a lot less effort.

Although things like RME are expensive, a used Arcam SR250 or SA30 gives you an all in one box with Dirac full bandwidth and good enough SINAD. Yamaha and Anthem have their integrateds with some room EQ. You have MiniDSP and of course, countless AV processors that have modern room EQ…
 

majingotan

Major Contributor
Forum Donor
Joined
Feb 13, 2018
Messages
1,560
Likes
1,861
Location
Laguna, Philippines
They fixed the Distortion Problem as you can see with the 90KHz bandwidth THD+N vs FR with the MIB hence the reason why I purchased this DAC.

1699144525494.png

thumbnail_IMG_8917.jpg
GetAttachmentThumbnail

I guess I have the second unit out of their production line based on the S/N :)
 
Last edited:

gvl

Major Contributor
Joined
Mar 16, 2018
Messages
3,528
Likes
4,129
Location
SoCal
Probably, as they said it solves "rising distorsion at high level & high frequencies"
which is exactly what you show

Not so sure as technically it’s not harmonic distortion.
 

majingotan

Major Contributor
Forum Donor
Joined
Feb 13, 2018
Messages
1,560
Likes
1,861
Location
Laguna, Philippines
Not so sure as technically it’s not harmonic distortion.

They reduced the distortion issue to undetectable levels by ear:

Original Yggdrasil More Is Less: Texas Instruments DAC11001A

1699196015147.png

More is better Texas Instruments DAC11001B

1699196142968.png


THD+N vs FR at 90 KHz bandwidth shows the big picture as well

index.php
 

CleanSound

Major Contributor
Forum Donor
Joined
Apr 30, 2023
Messages
1,654
Likes
2,533
Jason Stoddard wrote this; he is a legit EE. I thought DACs have no sound signature? I think the worse measuring Iggy was better then .05% THD+N, which is likely better than human hearing threshold. And he's talking about a double-blind test here too?

What's going on here @amirm ?
1699293027417.png
 
OP
G

GXAlan

Major Contributor
Forum Donor
Joined
Jan 15, 2020
Messages
4,038
Likes
6,302
Jason Stoddard wrote this; he is a legit EE. I thought DACs have no sound signature?

Transparency has no sound signature. We do have established loose and strict thresholds.



Recap of thresholds
Lenient

Dynamic range, linearity: 96 dB
THD, IMD: -66 dBFS / 0.05%
Noise: -85 dBFS / 0.005%
SINAD: 85 dB
Crosstalk: -60 dBFS
Jitter: -110 dBFS, -100 dBFS around the main tone
Frequency response: ±0.5 dB
Channel balance: 1 dB
Output impedance: 2 ohms

Strict
Dynamic range, linearity, SINAD: 120 dB
THD, IMD, noise, crosstalk, jitter: -120 dBFS / 0.0001%
Frequency response, channel balance: ±0.1 dB
Output impedance: 0.16 ohms



I think the worse measuring Iggy was better then .05% THD+N, which is likely better than human hearing threshold. And he's talking about a double-blind test here too?
0.05% THD is your lenient threshold and I bet many listeners would NOT be able to hear a difference but it’s not your strict threshold.

Until I see data, it’s also hard to know if there was a statistically significant difference or not.
 

NTK

Major Contributor
Forum Donor
Joined
Aug 11, 2019
Messages
2,779
Likes
6,174
Location
US East
They are not getting anywhere near the TI datasheet performance.

ti_dac11001b_1.png

ti_dac11001b_2.png
 

CleanSound

Major Contributor
Forum Donor
Joined
Apr 30, 2023
Messages
1,654
Likes
2,533
Transparency has no sound signature. We do have established loose and strict thresholds.



Recap of thresholds
Lenient

Dynamic range, linearity: 96 dB
THD, IMD: -66 dBFS / 0.05%
Noise: -85 dBFS / 0.005%
SINAD: 85 dB
Crosstalk: -60 dBFS
Jitter: -110 dBFS, -100 dBFS around the main tone
Frequency response: ±0.5 dB
Channel balance: 1 dB
Output impedance: 2 ohms

Strict
Dynamic range, linearity, SINAD: 120 dB
THD, IMD, noise, crosstalk, jitter: -120 dBFS / 0.0001%
Frequency response, channel balance: ±0.1 dB
Output impedance: 0.16 ohms




0.05% THD is your lenient threshold and I bet many listeners would NOT be able to hear a difference but it’s not your strict threshold.

Until I see data, it’s also hard to know if there was a statistically significant difference or not.
Reading what Jason is saying and having my limited knowledge of Schiit's design history, either he honestly swears by what he said, or he is a damn good snake oil peddler.

I just have a very hard time believer that he is this dirty of a snake oil peddler. I wish Jason can go on live streaming with a third party referee to observe this so called "double-blind" listening test with the participants he referenced.
 

AaronJ

Senior Member
Joined
Nov 9, 2021
Messages
325
Likes
583
Jason Stoddard wrote this; he is a legit EE. I thought DACs have no sound signature? I think the worse measuring Iggy was better then .05% THD+N, which is likely better than human hearing threshold. And he's talking about a double-blind test here too?

What's going on here @amirm ?
View attachment 324490
What’s crazy is he writes their most expensive product sounded broken compared to the new one. So their most expensive product was garbage? Shouldn’t everyone who owned the MIB be insulted? Is he trying to shame his loyal fan base into “upgrading” to the new chips?
 

NTK

Major Contributor
Forum Donor
Joined
Aug 11, 2019
Messages
2,779
Likes
6,174
Location
US East
But it must suppress 20khz harmonics if that’s the performance you’re referring to.
You notice that in the second plot in post #10, the biggest spike is ~30 kHz, and is not a harmonic of the fundamental of 14 kHz. H2 (28 kHz) is actually quite close to the datasheet. Even with a 30 kHz LP filter, the attenuation at the cutoff frequency would only be 3 dB. This is a rather less than perfect implementation of the chip.
 
Top Bottom